Which data subset should be augmented for deep learning? a simulation study using urothelial cell carcinoma histopathology images.

BACKGROUND: Applying deep learning to digital histopathology is hindered by the scarcity of manually annotated datasets. While data augmentation can ameliorate this obstacle, its methods are far from standardized. Our aim was to systematically explore the effects of skipping data augmentation; applying data augmentation to different subsets of the whole dataset (training set, validation set, test set, two of them, or all of them); and applying data augmentation at different time points (before, during, or after dividing the dataset into three subsets). Different combinations of the above possibilities resulted in 11 ways to apply augmentation. The literature contains no such comprehensive systematic comparison of these augmentation ways. RESULTS: Non-overlapping photographs of all tissues on 90 hematoxylin-and-eosin-stained urinary bladder slides were obtained. Then, they were manually classified as either inflammation (5948 images), urothelial cell carcinoma (5811 images), or invalid (3132 images; excluded). If done, augmentation was eight-fold by flipping and rotation. Four convolutional neural networks (Inception-v3, ResNet-101, GoogLeNet, and SqueezeNet), pre-trained on the ImageNet dataset, were fine-tuned to binary classify images of our dataset. This task was the benchmark for our experiments. Model testing performance was evaluated using accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve. Model validation accuracy was also estimated. The best testing performance was achieved when augmentation was done to the remaining data after test-set separation, but before division into training and validation sets. This leaked information between the training and the validation sets, as evidenced by the optimistic validation accuracy. However, this leakage did not cause the validation set to malfunction. Augmentation before test-set separation led to optimistic results. Test-set augmentation yielded more accurate evaluation metrics with less uncertainty. Inception-v3 had the best overall testing performance. CONCLUSIONS: In digital histopathology, augmentation should include both the test set (after its allocation), and the remaining combined training/validation set (before being split into separate training and validation sets). Future research should try to generalize our results.

A randomised trial comparing transurethral to percutaneous cystolithotripsy in boys.

OBJECTIVE: To compare the efficacy and morbidity of transurethral cystolithotripsy (TUCL) and percutaneous CL (PCCL) in the management of bladder stones in male children. PATIENTS AND METHODS: A total of 100 boys, aged <14 years with a single bladder or urethral stone of <30 mm, were randomised into two equal groups. Initial diagnostic urethro-cystoscopy and push back of urethral stones were done for patients in both groups. Patients in Group A had TUCL, while those in Group B had PCCL through a 20-F sheath using a 12-F nephroscope. The two groups were compared regarding preoperative criteria, intraoperative details, and postoperative outcomes. RESULTS: The patients in this study had a median (range) age of 36 (4-144) months and stone size of 10 (5-26) mm. There was no statistically significant difference between the two groups for preoperative criteria. The assigned procedure was successful in 48 (96%) patients in Group A and 49 (98%) in Group B (P = 1). Complications were encountered in 11 (22%) patients in Group A and five (10%) in Group B (P = 0.171). The median (range) operative time was 21.5 (4-90) min in Group A and 13 (5-70) min in Group B (P < 0.001). In all, 47 (94%) stones needed disintegration in Group A vs 22 (44%) in Group B (P < 0.001). CONCLUSION: Both techniques have comparable success and complications rates. However, PCCL has a shorter operative time and less need for stone disintegration.

Critical Application of Adult Nephrolithometric Scoring Systems to Children Undergoing Mini-Percutaneous Nephrolithotomy.

Objective: To evaluate and compare the ability of the Guy's stone score (GSS), the S.T.O.N.E. nephrolithometry, and the Clinical Research Office of the Endourology Society (CROES) nomogram to predict the outcome of mini-percutaneous nephrolithotomy (MPNL) in children, and to identify which of the predictors involved in these scoring systems can separately affect this outcome. Patients and Methods: All children younger than 14 years who had MPNL in our center over a period of 3 years were included prospectively. Bivariate analyses were done to evaluate the associations of the three scoring systems and the predictors composing them with single-session stone clearance and complications. Receiver operating characteristic (ROC) curve analyses of the three scoring systems were conducted to evaluate and compare their abilities to predict the outcomes. Decision curve analyses for the three scoring systems were conducted to evaluate the clinical benefit of using each of them to predict stone clearance. Results: We consecutively enrolled 92 renal units in 89 children with a median age of 9.5 years. Single-session stone clearance was achieved in 76 (82.6%) renal units. Complications occurred with 19 (20.7%) procedures. Stone multiplicity (p = 0.043), staghorn stone (p = 0.007), prior stone treatment (p < 0.001), number of calices involved (p = 0.006), stone burden (p = 0.003), GSS (p < 0.001), S.T.O.N.E. nephrolithometry (p = 0.012), and CROES nomogram (p < 0.001) had significant associations with stone clearance. Only stone attenuation was significantly associated with complications (p = 0.031). For prediction of stone clearance, CROES nomogram demonstrated the greatest area under the ROC curve and the greatest net benefit on decision curve analyses. Conclusions: For children undergoing MPNL, CROES nomogram is the best to predict stone clearance. However, none of the studied scoring systems predicted complications efficiently.

Health related quality of life after urinary diversion. Which technique is better?

OBJECTIVE: To compare quality of life (QoL) after urinary diversion (UD) following radical cystectomy (RC) using validated questionnaires. PATIENTS AND METHODS: Between January 2011 and June 2016, 150 patients (121 men [80.7%] and 29 women [19.3%]) with invasive bladder cancer who underwent RC and UD were included in this prospective study. Patients were divided into 2 groups; group I included the orthotopic neobladder 50 (33.3%) and uretro-sigmoidostomy 41 (27.3%) and group II included uretero-cutanoustomy 33 (22.1%) and ileal conduit 26 (17.3%) patients. QOL was evaluated using the Functional Assessment of Cancer Therapy-Bladder Cancer. The erectile function (EF) was assessed using the Sexual Health Inventory for Men Questionnaire. Evaluation was done before and after one year postoperatively. RESULTS: The mean ±â€¯SD patient age was 55.0 ±â€¯7.9 and 59.5 ±â€¯8.5 years in both groups, respectively (p = 0.001). There was a significant difference in the physical, social/family, emotional and functional statuses that were significantly higher in group I. One year postoperatively, the emotional well-being became insignificantly different, but other QoL parameters remained significantly different between both groups. Regarding EF, there was a significant difference between patients who underwent nerve-sparing (No. 29) and non nerve-sparing RC (No. 59) (p < 0.001). CONCLUSIONS: Which type of diversion is the best is still a controversial topic. Egyptian patients may prefer the continent UD to avoid the urostomy appliance and its associated daily-life constraints. Detailed patient counseling and active participation of the patient in selecting the treatment methods are important for better postoperative QoL.

Latent tuberculosis infection among patients with erectile dysfunction.

No previous studies have investigated the prevalence of latent tuberculosis infection (LTBI) among patients with erectile dysfunction (ED) or its contribution to the development of high-grade ED through a process of chronic inflammation-induced atherosclerosis. The aim of this study was to determine the frequency of LTBI among patients with erectile dysfunction and to explore the contribution of LTBI to high-grade ED. For all the study sample, clinical evaluation, imaging studies, and laboratory investigations were provided. Evaluation included, but was not confined to, scrotal ultrasonography, tuberculin skin test, and QuantiFERON-TB Gold test. The study sample mean ± SD age was 47.9 ± 13.6 years. Approximately 30% of the patients had LTBI and 43% had high-grade ED. After a multivariate analysis, it was found that older age (≥40 years) (OR, 5.2; 95% CI, 1.9-54.6; p 0.004), metabolic syndrome (MS) (OR, 3.4; 95% CI, 1.3-48.2; p 0.016), and LTBI (OR, 4.1; 95% CI, 1.7-61.3; p 0.021) were significantly, independently associated with high-grade ED as opposed to low-grade ED. In conclusion, the prevalence of LTBI among patients with high-grade ED is higher than among those with low-grade ED. In addition to LTBI, older age and MS are associated with high-grade ED as opposed to low-grade ED.

Is Concealed Epispadias a Rare Variant?

OBJECTIVE: To present our experience with concealed epispadias and to estimate its actual share in the isolated male epispadias cases and its effect on the surgical outcome. MATERIALS AND METHODS: Consecutive patients with isolated male epispadias treated in our center between 2008 and 2015 were classified into concealed and classic epispadias. The 2 groups were compared regarding age at presentation, meatal location, incontinence, dorsal curvature, success rate, and complications. RESULTS: Out of 51 patients with isolated male epispadias, 11 (21.6%) were concealed: 7 balanic and 4 penile shaft epispadias. Concealed epispadias cases were found to have significantly delayed age at presentation, more distal meatal location, and less incontinence rate than classic epispadias cases. None of the surgical outcome parameters showed significant difference between the 2 groups. CONCLUSION: Concealed epispadias represents about one-fifth of isolated male epispadias cases. Impediment and delay of diagnosis are its main clinical impacts, with insignificant effect on the surgical outcome.